Anthropic staff expect far faster AI progress
An Anthropic survey finds its staff and the US public mostly agree on what AI can do today but split sharply on future progress: 91% of staff, versus 39% of adults, expect Nobel-level AI discoveries by 2030.
Anthropic has published survey research comparing how its own employees and a nationally representative sample of US adults expect AI to develop. The authors are Szymon Sacher, Tess Cotter, Charles I. Jones, Anton Korinek and Peter McCrory, and the surveys ran in August 2026.
Both groups largely agree on what AI can do now and rank eight tasks in nearly the same order of difficulty. The tasks run from writing routine emails to making a Nobel-level discovery. Their views split on what comes next. About a quarter of each group say AI can already design, conduct and publish scientific studies. By next year, nine in ten staff expect it to manage that, compared with 38% of the public. Between 19% and 40% of US adults say AI will never be able to do a given task. No more than 2% of staff say the same.
The gap is widest when people consider their own jobs. On average, staff expect AI to be able to do 75% of their current work by 2030, against 28% for US workers. Staff also expect faster adoption, with AI used for 65% of the tasks it could do by 2030, compared with 42% in the public's estimate. US adults named customers' preference for dealing with a person as the biggest obstacle to adoption.
Both groups reason about jobs in the same way, though. In each group, people who expect AI to take over more of their own work forecast higher unemployment, at nearly the same rate.
The findings measure opinion, not capability. One of the two groups surveyed is Anthropic's own staff.